# T2I Metric Evaluation on CROCsyn Inverse Image-Based Source: [T2I Metric Evaluation on CROCsyn Inverse Image-Based]() Collection index: [llms.txt]() Data updated: 2026-04-21T10:45:48.352Z Evaluation of Text-to-Image (T2I) metrics, specifically using the Scaled Accuracy score, on the CROCsyn Inverse Image-Based dataset. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | T2I Metric Evaluation | | Dataset (including any stated split) | CROCsyn Inverse Image-Based | | Metric | Scaled Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 18 of 18 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Scaled Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Scaled Accuracy | Result date | Report paper | | --- | --- | --- | --- | | VQAScore | 98.4 | Not recorded | 68fa789f1e209131cea796a5 | | AlignScore | 97.9 | Not recorded | 68fa789f1e209131cea796a5 | | BLIP2-ITM | 96.5 | Not recorded | 68fa789f1e209131cea796a5 | | BVQA | 94.9 | Not recorded | 68fa789f1e209131cea796a5 | | PickScore | 91.8 | Not recorded | 68fa789f1e209131cea796a5 | | VQAScore | 73.3 | Not recorded | 68fa789f1e209131cea796a5 | | BLIP2-ITM | 68.4 | Not recorded | 68fa789f1e209131cea796a5 | | AlignScore | 65.9 | Not recorded | 68fa789f1e209131cea796a5 | | BLIP2-ITM | 65.7 | Not recorded | 68fa789f1e209131cea796a5 | | VQAScore | 64.2 | Not recorded | 68fa789f1e209131cea796a5 | | AlignScore | 62.2 | Not recorded | 68fa789f1e209131cea796a5 | | BVQA | 60.3 | Not recorded | 68fa789f1e209131cea796a5 | | CLIPScore | 58.9 | Not recorded | 68fa789f1e209131cea796a5 | | PickScore | 56.4 | Not recorded | 68fa789f1e209131cea796a5 | | BVQA | 55.4 | Not recorded | 68fa789f1e209131cea796a5 | | PickScore | 52.9 | Not recorded | 68fa789f1e209131cea796a5 | | CLIPScore | 37.2 | Not recorded | 68fa789f1e209131cea796a5 | | CLIPScore | 30.4 | Not recorded | 68fa789f1e209131cea796a5 | ### VQAScore Scaled Accuracy: 98.4. Result ID: 69e754d647f34ce924ce5b10. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"entity placement (EP)","Metric Category":"VQA-based"}. This record does not assert independent reproduction. ### AlignScore Scaled Accuracy: 97.9. Result ID: 69e754d647f34ce924ce5b0b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"entity placement (EP)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. ### BLIP2-ITM Scaled Accuracy: 96.5. Result ID: 69e754d647f34ce924ce5b0c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Prompt Type (P)":"entity placement (EP)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. ### BVQA Scaled Accuracy: 94.9. Result ID: 69e754d647f34ce924ce5b0f. [Reporting paper]() Recorded evaluation configuration: {"Prompt Type (P)":"entity placement (EP)","Metric Category":"VQA-based"}. This record does not assert independent reproduction. ### PickScore Scaled Accuracy: 91.8. Result ID: 69e754d647f34ce924ce5b0e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Prompt Type (P)":"entity placement (EP)","Metric Category":"Tuned"}. This record does not assert independent reproduction. ### VQAScore Scaled Accuracy: 73.3. Result ID: 69e754d647f34ce924ce5b16. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"property variation (PV)","Metric Category":"VQA-based"}. This record does not assert independent reproduction. ### BLIP2-ITM Scaled Accuracy: 68.4. Result ID: 69e754d647f34ce924ce5b12. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Prompt Type (P)":"property variation (PV)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. ### AlignScore Scaled Accuracy: 65.9. Result ID: 69e754d647f34ce924ce5b11. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"property variation (PV)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. ### BLIP2-ITM Scaled Accuracy: 65.7. Result ID: 69e754d647f34ce924ce5b06. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Prompt Type (P)":"entity variation (EV)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. ### VQAScore Scaled Accuracy: 64.2. Result ID: 69e754d647f34ce924ce5b0a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"entity variation (EV)","Metric Category":"VQA-based"}. This record does not assert independent reproduction. ### AlignScore Scaled Accuracy: 62.2. Result ID: 69e754d647f34ce924ce5b05. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"entity variation (EV)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. ### BVQA Scaled Accuracy: 60.3. Result ID: 69e754d647f34ce924ce5b15. [Reporting paper]() Recorded evaluation configuration: {"Prompt Type (P)":"property variation (PV)","Metric Category":"VQA-based"}. This record does not assert independent reproduction. ### CLIPScore Scaled Accuracy: 58.9. Result ID: 69e754d647f34ce924ce5b0d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"entity placement (EP)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. ### PickScore Scaled Accuracy: 56.4. Result ID: 69e754d647f34ce924ce5b14. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Prompt Type (P)":"property variation (PV)","Metric Category":"Tuned"}. This record does not assert independent reproduction. ### BVQA Scaled Accuracy: 55.4. Result ID: 69e754d647f34ce924ce5b09. [Reporting paper]() Recorded evaluation configuration: {"Prompt Type (P)":"entity variation (EV)","Metric Category":"VQA-based"}. This record does not assert independent reproduction. ### PickScore Scaled Accuracy: 52.9. Result ID: 69e754d647f34ce924ce5b08. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Prompt Type (P)":"entity variation (EV)","Metric Category":"Tuned"}. This record does not assert independent reproduction. ### CLIPScore Scaled Accuracy: 37.2. Result ID: 69e754d647f34ce924ce5b13. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"property variation (PV)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. ### CLIPScore Scaled Accuracy: 30.4. Result ID: 69e754d647f34ce924ce5b07. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt Type (P)":"entity variation (EV)","Metric Category":"Embedding-based"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.